If you’re got a home-grown Elasticsearch environment, then you might not be thinking about backup automation. You might be relying on a very primitive backup solution, one that doesn’t offer any of our advanced features. If you’re like we were back in the day, you’ve got some rough measure of data security—at the cost of flexibility and periodic lapses in cluster responsiveness.
Although we’ve had backup and restore capability for quite sometime now, we’re realizing that it’s good to make a followup announcement. We have an automatic daily backup feature, which is built around the snapshot/restore API from Elasticsearch. Take a moment to read through this article and consider your options on backup and restore.
Since we fully exploit the snapshot/restore API, we here at Qbox have wide flexibility to accommodate varying requirements of our customers and their cluster configurations. Most importantly, we precisely tune the snapshot process to prevent any adverse impact to user operations.
All Qbox clusters get a daily snapshot, and we store all files in a remote, redundant repository for at least 7 days. In our cluster dashboard, you can easily perform a manual restore for any cluster to any state within the previous week.
It’s important that we make note of one Elasticsearch restriction: during the execution of a snapshot, no other snapshots can begin until that process is complete.
One more thing: the new backup system is available only for Qbox clusters running Elasticsearch v1.1.1 and later. If you happen to be running an older version and need to enable backups, please send a request to email@example.com and ask us for a prompt upgrade to the latest version of Elasticsearch.